An optical water property important for the fauna and - in coastal regions - for human beings, is the in-water visibility, which can be expressed quantitatively in terms of the Secchi depth (ZSD, §5.1). Chen et al. (2007) investigated the relationship between in situ measured Secchi depth and SeaWiFS radiances in Tampa Bay, Florida. For the 1997 to 2005 period, 80 matching data pairs were recovered. An empirical relationship
was found between Secchi depth and the value of Kd(490) calculated from the SeaWiFS radiances using the semi-analytical algorithm of Lee et al. (2005) (see above): the authors suggest that this could be used for routine monitoring of this aspect of water quality in coastal and large estuarine waters such as Tampa Bay.
For the Peel-Harvey estuary in Western Australia, Lavery et al. (1993) studied the relationship between radiance values in Landsat TM Bands 1 (450-520 nm) and 3 (630-690 nm), atmospherically corrected by the dark-pixel method, and Secchi depth. Using the algorithm
Zsd = 0.74 - 0.05Lw(630 - 690) + 1.80Lw(45° - ^ (r2 = 0.81)
they mapped the distribution of ZSD in this estuarine system. On the basis of in situ measurements in a number of lakes in central Spain, Dominguez Gomez et al. (2009) arrived at an empirical relationship between Secchi depth and above-surface reflectance in the 520 to 600 nm waveband (corresponding to TM Band 2)
With the help of atmospherically corrected Landsat TM images, this relationship was then used to monitor water transparency in lakes in the study area at intervals over the period 1984 to 2000.
We saw earlier that ZSD is approximately proportional to 1/(c + Kd), so if (c + Kd)-1 can be estimated from remotely sensed radiances then in-water visibility could be mapped. Doron et al. (2007) have developed an algorithm to achieve this, using radiance values only in two wavebands, 490 and 709 nm. At 490 nm, absorption and scattering by all seawater constituents contribute to variability in reflectance, whereas at 709 nm -where water absorbs strongly but other components hardly at all -variations in reflectance are almost solely due to changes in the particle backscattering coefficient. The absorption coefficient at 709 nm (a) is assumed to be equal to that of pure water, which is known. The constant of proportionality f) relating irradiance reflectance to bb/a in eqn 6.5 is, on the basis of literature data, given the value 0.335, for both wavebands. From reflectance at 709 nm, bb(709) is thus obtained. Using an assumed ratio between particulate backscattering at 490 nm and that at 709 nm, the backscattering coefficient (bbp) at 490 nm, due to particles, is obtained. From bbp(490), the total particulate scattering coefficient (bp) is obtained with the help of an empirical relationship expressing the total particulate scattering coefficient (bp) as a function of the particulate backscattering coefficient, at 490 nm, derived on the basis of a substantial data set obtained from in situ measurements at a large number of stations in Case 1 and Case 2 ocean waters. Addition of scattering due to water (known) to that due to particles gives b(490), the total scattering coefficient. The absorption coefficient at 490 nm, a(490), is obtained from bb(490) and reflectance at 490 nm, using eqn 6.5. The beam attenuation coefficient is then obtained from c(490) = a(490) + b(490). Using the Lee et al. (2005) algorithm, Kd(490) can be estimated from reflectance data, so that [c(490) + Kd(490)]-1 can finally be arrived at. Values of [c(490) + Kd(490)]-1 calculated from in situ reflectances as described, correlated closely (r2 = 0.90) with those obtained by direct measurement of c(490) and Kd(490) within the water.
Photosynthesis in the aquatic environment
Was this article helpful?